Predicting arterial oxygen tension from maximum expiratory flow volume curves in cystic fibrosis

Measurements of arterial oxygen tension (PaO2) while breathing room air, and maximum expiratory flow volume curves were performed in 34 patients with cystic fibrosis (age range 7-27 years, 24 males and 10 females). Logistic regression was performed using forced vital capacity (FVC), forced expiratory volume in 1 second (FEV1), peak expiratory flow rate (PEFR), and forced expiratory flow at 75% of expired vital capacity (FEF75) to model an equation for predicting when PaO2 would be less than 55 mmHg (severe hypoxemia). Equations were modelled using one, two, three, or all four of the variables. For the univariate logistic regression, each of the four variables was a significant (P less than 0.001) predictor for severe hypoxemia. FVC was the best predictor with an R2 = 0.56, sensitivity of 100% (false negative rate = 0%), and a specificity of 88.5% (false positive rate = 27%). The model predicted that patients with an FVC less than 35% of the predicted normal were at risk of having PaO2 less than or equal to 55 mmHg. Adding FEV1, FEF75, or PEFR in various combinations to FVC made the model equation more complicated but did not add significantly to the ability to predict severe hypoxemia.